Distributors benefit from e-commerce platforms that deliver better customer experiences to their buyers, but there’s often a gap between what buyers want and what businesses deliver.

Digital experience platforms (DXP) can help even the slowest digital adopters boost their e-commerce presence, growth, digital revenue and improve profitability, according to MDM’s recent webcast, “How a DXP Can Boost Your E-Commerce Sales.”

“When I’m talking about a DXP, what I’m talking about is the platform that essentially helps you manage one of the biggest challenges for B2B e-commerce, which is, what is the experience that your customers are having with the information on the product lines that you’re showing?” said Jason Hein, principal B2B strategist at Bloomreach, during the webcast.

Distributors’ customers want websites that work like their outside sales staff. This is where site search recommendations come into play. But, if a website is poorly designed, it may take scrolling through numerous pages before customers find the product they are looking to purchase. According to research cited by Hein, 25% of B2B business decision makers said direct websites are their fastest-growing channel, while 56% of B2B customers would pay more for a better or faster buying experience.

While DXP-enabled e-commerce platforms can provide a better customer experience for existing customers and reel in new ones, 50% of B2B customers are unlikely to purchase from the same seller again if they had a bad experience, said Hein.

“If that experience was really terrible — you’re not recognizing the terms that I’m using and the keywords and category structure are terrible — the recommendations aren’t meaningful,” Hein said. “How do you know your [customers’ web] experience is bad? Well, it’s pretty easy. Generally, your customers will tell you. They’ll be complaining, ‘Oh, your search doesn’t work. You’re trying to sell me something that I don’t need anymore,’ or, ‘that products is not going to work for what I need.’”

By using machine learning and artificial intelligence, DXP e-commerce platforms can learn what categories, description and keyword searches are the most meaningful for a distributor’s website.

“What a DXP does is it essentially is a tool that tracks the behavior of every user who comes to your site and analyzes it,” Hein explained. “And it actually looks at: What is that customer doing? What pages are they looking at? What queries are they searching for? What recommendations and widgets are they clicking on? What results are they clicking on?”

Hein said that while DPX capabilities are reminiscent of data analytics, the difference is that DXP gives distributors the ability to use machine learning tools to scale data into reports and automatically make decisions about criteria such as customer searches.

“I’m showing that customers who searched for this term bought this product, but then also bought this other product,” Hein said. “So just think of the ability to have every single one of those transactions rolled up into an algorithm that can make merchandising decisions for you. What are customers looking for? How do customers decide what to view versus what to purchase? How can I get more customers exposed to more of my selection?

“This is actually one of the big challenges that Amazon is going to face. Amazon has a massive assortment of [roughly] 75-billion SKUs, because that’s been a part of their driver — this endless-aisle concept. But the problem is, if you don’t have a discovery experience that can get you to all of those products, then there’s really no point in having so many SKUs.”

Bloomreach, which counts Watsco, HD Supply, Sonepar and Global Industrial among its distributor customers, has built machine-learning datasets that are specific to e-commerce while also using algorithms to help optimize that data increase revenues.

AI is not the rise of machines

One of the common misconceptions about artificial intelligence (HBO’s “Westworld” or the movie “Ex Machina,” for example) is that artificial intelligence-based platforms will eventually run amok. During the webcast, one of the attendees asked if AI could be trusted for the merchandising of technical products, or whether a distributor would still have control over the algorithms. According to Hein, the answer to both questions was, “Yes.”

“Here’s the great thing about AI,” Hein said. “AI is thought of as this kind of scary thing. But fundamentally, if AI is truly artificial, it’s not going to be intelligent, and if it’s really intelligent, that means it’s not artificial. Because we’re using a machine learning component, the recommendations that the algorithm starts to generate are based not on some sort of speculative thing, it’s really based on the sum total of all of your customers’ behaviors. So what gets served up is not a speculation of the algorithm, it’s really a distillation of what your customers are doing.

“We have customers who basically let the algorithm handle everything, and we have customers who have also said, ‘There are certain things we want to be able to tune in.’ When it comes to enterprise software, there’s always some degree of tweaking that happens, and that’s one of the conversations that we have with customers.”

To listen to the webcast in its entirety, click here.